Myth #2: Accurate Forecasts are Achieved by Scrubbing Data and Using Weighted Averages

By Michael Lock

Sales organizations spend an immense amount of time scrubbing data. Usually, an army of sales managers spend time looking at deals, pipelines and forecasts. A lot of time is spent defining key deal metrics such as sales stage, forecast category, deal amount, and industry classification. An even larger amount of time is spent defining customer data fields in CRM in hopes that we might influence deal execution and closure rates. Then, we pay an army of sales trainers, sales ops people and middle managers to scrub that data constantly so we can get the deals in the right “buckets”. Next, we take a simple math formula and multiply the amount of deals in each bucket by some mythical close rate to produce the forecast. Then we apply a bit of old-school know-how. I often see sales leaders submit a forecast with 90% of their pipeline in the “commit” bucket, 50% in their “most likely” bucket, and 10% of their “pipeline” bucket–a.k.a. the 90-50-10 rule.

The 90-50-10 rule doesn’t work. It may work in the final week of the quarter, but it is wildly inaccurate at almost every other time. We have all heard the stats that 50% of deals in commit never close, and we should all be aware that forecasting with the weighted average pipeline is as outdated as listening to music with an eight-track tape.

I am not saying that getting clean data is a bad idea. It’s a good idea! But spending massive amounts of time scrubbing data is a huge time-waster.  We will never get it scrubbed well enough, and even if we do, business is complex and fast changing so that applying simple math to forecast buckets does not work for two reasons.

First: Business is more complex than the simple forecast buckets that you have set up. An “upsell” deal in the “most likely” bucket has a higher probability to close than a net-new business deal in the “most likely” bucket.  A deal in “commit” with the best product will close at a much higher rate than with a less competitive product in the same bucket. Deals in different business segments or industries often behave differently. Even of we scrub our data correctly, we are not likely to be able to accurately forecast these complexities by putting them in three buckets.

Second: The bias of human judgement gets in the way. Every person has a different view of what the words “probability to close” means.  A deal that one rep judges as 50% probability to close, might be judged by another rep at 90%. What one rep classifies as “commit” might be judged by another rep as “upside”. To complicate matters further, we are not dealing with one level of judgment. We have to consider that the reps’ judgment is “augmented” by the sales manager’s judgment and then “augmented” again by the judgement of the VP. With all this judgement, the chances of a deal being accurately reflected in a multi-level sales organization are slim. And Slim is out of town.

There is a solution. The answer is data science and AI.   On Wall Street, quants have taken over. We need more quants in sales.  Like Google, Amazon and Netflix, who all use AI to showcase programming, products or ads that are tailored to our individual interests.  Today’s sales leaders need data, proven algorithms and AI to see what is truly going on in their business.  The days of 90-50-10 are over.

See how Aviso can help